The Internet is made up of autonomous systems. Each autonomous system is a network operated by a single organization, and has routing connections to some number of neighboring autonomous systems.
This image depicts a graph of 19,869 autonomous system nodes joined by 44,344 connections. The sizing and layout of the autonomous systems is based on their eigenvector centrality, which is a measure of how central to the network each autonomous system is—an autonomous system is central if it is connected to other autonomous systems that are central. This is the same graph-theoretical concept that forms the basis of Google’s PageRank algorithm.
The graph layout begins with the most central nodes and proceeds to the least, positioning them on a grid that subdivides after each order of magnitude of centrality. Within the constraints of the current subdivision level, nodes are placed as near as possible to previously placed nodes that they are connected to.